sammccall added a comment.

As discussed offline, basically the only thing to do for testing ThreadPool is 
to bash it with a workload, and some sort of whole-program stress test seems 
ideal for this and will also give some coverage to other components (and we 
should run under tsan!).

TUScheduler on the other hand is a big important class with a clear interface 
and contract, and is about to get a new implementation - testcases verifying 
the contracts will be extremely valuable.
These tests don't really need to be concurrency-heavy I think.



================
Comment at: clangd/TUScheduler.h:10
+
+#ifndef LLVM_CLANG_TOOLS_EXTRA_CLANGD_THREADING_SCHEDULER_H
+#define LLVM_CLANG_TOOLS_EXTRA_CLANGD_THREADING_SCHEDULER_H
----------------
header guards are stale, sorry!


================
Comment at: clangd/Threading.h:10
+
+#ifndef LLVM_CLANG_TOOLS_EXTRA_CLANGD_THREADING_THREADPOOL_H
+#define LLVM_CLANG_TOOLS_EXTRA_CLANGD_THREADING_THREADPOOL_H
----------------
this one also stale


Repository:
  rCTE Clang Tools Extra

https://reviews.llvm.org/D42174



_______________________________________________
cfe-commits mailing list
cfe-commits@lists.llvm.org
http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

Reply via email to